Why didn't the US and USSR jam each other's early-warning radar?

Unless they were actively planning on launching an attack (or anticipating an imminent "hot war"), there was no real reason to do so. In reality, given how destructive such a war would be, both sides had a strong incentive to avoid outright war with each other.

Another answer (correctly, in my opinion) pointed out the very real possibility that such an action would cause the other side to assume that there was an ongoing (or imminent) attack, so this would risk the other side retaliating by launching missiles. Even if the other side didn't retaliate with missiles, they would almost certainly retaliate in some other way. At a minimum, they'd probably try to do the same thing to the other side, which would in turn leave them equally vulnerable to an attack (possibly even by some hostile third party who might seize the opportunity to frame the other superpower for the attack).

By way of example, think about the Cuban Missile Crisis: what a lot of people forget about that is that placing the missiles there was in retaliation for the U.S. putting missiles in Turkey.

The point being, something that would result in a minor theoretical advantage in a hot war that neither side really wanted anyway would almost certainly result in concrete retaliation (if not outright war).

How will they jam each other's radars without getting detected? They don't share a land border. No jammer is powerful enough to jam a radar across an ocean. The jammer will have to be placed in a ship and will give away the location once it starts operating.

A friend owned a medium-large dog, a male. We went for a walk one day, following a fence between her yard and her neighbor's, also between her dog and her neighbors male dog. As soon as they saw each other, they growled, bared their teeth, ears back, eyes wide open, frothing. They carried on this way until we reached a gap in the fence. Both dogs gave a start, became mute and immediately backed up so that the fence was between them again and commenced threatening each other with assured destruction.

That, in a nutshell, was the cold war. There had to be a fence that prevented the US and the USSR from killing each other. I wouldn't be surprised if we were sharing each other's radar technologies, in secret of course, just to be sure both had the latest and best and neither could be absolutely sure that their missiles wouldn't get past the other's defenses.

Radar jamming is not trivial

First of all, early warning radar (EWR) operate at very low frequencies, for example 7-19 MHz for Duga. For comparison, usual low frequency acquisition radars like P-19 operate in UHF band (300 MHz - 3 GHz). This is the reason why they need such large receivers and transmitters (rule of thumb, lower the frequency, bigger the antenna). Usual jamming aircraft would have difficulty even determining have they been "painted" by low frequency radar, not to mention the part where they had to transmit at the same frequency. Also, jamming techniques could be reduced to two groups:

• Brute force jamming where you simply emit stronger signal at same frequency towards receiver. To do this, you would need to know emitting frequency of the radar (not easy, because as we mentioned before, it is not easy to detect low frequency radar) which might do "frequency hopping" while emitting. Also, your jamming platform is bound to be discovered, and likely destroyed if it emits for prolonged time interval - it is quite simple to triangulate the position of a jammer. Another concern is that during the arms race both sides developed various anti-jamming techniques like filtering signal from certain azimuth, aforementioned frequency hopping, or even shaping the signal to avoid random patterns from the jammers.

• False targets is much more devious technique where you would attempt to create decoys for targeted radar system, either electronically or mechanically. Electronic false targets are usually created by jamming aircraft and ships, but unlike brute force jamming, they do not attempt to blind the radar, instead creating fake returns that resemble real targets. Again, because of low frequency, doing something like that is not easy. Mechanical decoys are smaller and cheaper than real radar targets (in this case nuclear missiles), but through various reflective techniques they do have similar radar return. One thing that is hard to mimic is missile speed and height. Usually, decoys would be deployed from the nuclear missile itself to overcome this.

Anyway, any successful jamming technique had to be kept a secret - otherwise the opponent could develop counter-measures or simply be frightened enough to launch first. The downside of this is that you could never be sure that your jamming actually works in the real world. Unlike conventional weapons, there were no proxy wars to test nuclear weapons.

The short answer is "Mutual Assured Destruction." If you jam the other side's radar effectively, the natural assumption is that you're attacking, so they'll attack in response. So you will set off a nuclear war with your first serious jamming attempt.

Under normal circumstances, you want the other side's early-warning radar to be working well so that they can see you are not attacking. There was at least one incident when a nuclear war nearly started accidentally owing to an early-warning system (not radar) malfunction.

The only time jamming makes sense is when you've already launched an all-out surprise attack. You start jamming just after the point when you believe the enemy will be aware of the scale of your attack. That handicaps their ability to take defensive precautions, a bit. But that's the most use it can be. Overall, a comprehensive, and thus expensive, jamming system is unlikely to be worth its cost.